Hi Raj.

you can have a good start with this link

http://statwiki.kolobkreations.com/wiki/Main_Page

Hey,

try this one:

http://ncme.org/linkservid/47EFEB5A-1320-5CAE-6EC90BFDF09AA39E/showMeta/0/

In addition to the resources listed above, you might look at Madhu Viswanathan's Measurement Error and Research Design (link is to to Kindle edition, which can be rented).

It was a great help to me in my dissertation (recommended to me by my chair, Dr. Amit Ghosh).

http://www.amazon.com/Measurement-Error-Research-Design-Viswanathan-ebook/dp/B004BA6QAI/ref=tmm_kin_title_0?_encoding=UTF8&sr=8-1&qid=1413647791

An excellent resource is John Loehlin's Latent Variable Models (2004) . This is one of the best and most readable resources out there.

Also, for an excellent SEM package, use the lavaan package in R (for more information about lavaan go to http://lavaan.ugent.be  .  Then work through the examples there.

lavaan is open source and very powerful.

Bill

The conditions vary, depending for example of the measurement level of the variables. For example, maximum likelihood estimation is for interval (or ratio) measurement level variables that are suposedd to be sampled form a multivariate normal distribution. But of couse you have other estimation method for ordinal level variables.

A good introduction could be fund here: http://ncme.org/linkservid/47EFEB5A-1320-5CAE-6EC90BFDF09AA39E/showMeta/0/